The 2014 licensing deal with Merlin and Pandora played a surprising role in setting 2016 statutory webcasting rates.

Two licensing agreements — one by a group of independent labels and another by a major label — played leading roles in determining what Pandora, iHeartRadio and other webcasters (think “internet radio”) will pay for the music they play over the next five years. The independent labels’ agreement provided the biggest surprise and helped create rates that were beneficial to webcasters and detrimental to rights holders (like those indie labels) and performing artists.

The proceeding on webcasting rates for 2016 through 2020, called Webcasting IV, was a contest to influence what benchmarks the three Copyright Royalty Board judges would employ to set new statutory webcasting rates. These royalties are paid by services to SoundExchange, which then distributes that money to labels and performing artists. The public release of the full, although redacted, 203-page determination provides insight into how the judge’s thinking about the evidence provided to them.

At its core, the rate proceeding isn’t difficult to understand. The various parties — rights owners like labels on one side, webcasters on the other — present evidence and testimony in proposing benchmarks that the judges will use to set the statutory rates for free and subscription non-interactive services like Pandora. These parties attempt to convince the judges to adjust those benchmarks either upward, in the case of rights owners, or downward, in the case of webcasters. “If a benchmark is deficient in some manner, the adversarial process of this proceeding allows the parties to expose those deficiencies,” the judges wrote.

But it’s more complicated than that; a rate proceeding is not like a market transaction. The judges attempt to set “reasonable rates” that reflect what a willing buyer and willing seller would agree to in a hypothetical competitive market. As the judges noted (citing Section 114(f)(2)(b) and the D.C. Circuit Court) the best place to look is rates that have already been negotiated in an actual, competitive market. In fact, the proceeding was “replete with evidence that the parties entered into various transactions with the knowledge, if not the intent, that such agreements could be used as evidentiary benchmarks in this proceeding,” the judges wrote.

The judges gamely called the divide between those two previously mentioned rates (between Merlin and Pandora, and between Warner Music Group and iHeartRadio) their “zone of reasonableness” for setting a new free-to-play, non-interactive rate. The lower reach of that zone was set by a 2014 licensing agreement between independent rights organization Merlin and the largest webcaster, Pandora. The upper, a 2013 agreement between Warner Music Group and iHeartRadio (the latter was still Clear Channel at the time). No other deals, such as Big Machine’s 2012 agreement with iHeartRadio, came into play. The judges would split the difference between the two benchmarks, according to indie and major market shares, and arrive at a rate for free services of 0.17 cents. A separate benchmark was set for subscription non-interactive services (such as Pandora One).

The Merlin/Pandora agreement that the judges examined features two types of rates; a headline rate for a “natural” level of streaming quantity of plays and a lower, “steered” rate for incremental plays that takes advantage of Pandora’s ability to “modify its playlist-selecting algorithms to rely more or less heavily on the music of a particular record company,” as the judges described it. Merlin opted to trade extra streams for a lower royalty rate. The end result would be a lower average rate but a greater amount of royalties (money).

That agreement came under fire immediately after that rate was revealed. Critics worried — correctly, as it would turn out — the CRB’s judges might be influenced by the presence of the lower, steered rate. If the judges set 2016 rates based on an average per-play rate, not the headline rate negotiated by Merlin, all labels would be negatively affected.

Even the slightest movement in the statutory rate can have a large economic impact. Just as an example, and using hypothetical revenue figures, imagine webcasters will pay SoundExchange $600 million in statutory royalties in 2016 (a possible scenario). The difference between the actual 2016 rate of 0.17 cents and the next possible increment, 0.18 cents, would be $60 million. Over the following four years, the impact of that tenth of a cent would grow into the hundreds of millions of dollars. (Again, a hypothetical example.)

But where to set the benchmarks? Pandora believed the steered rate in its Merlin agreement, from 0.1105 to 0.1205 cents depending on the amount of steering, should be the benchmark. It argued its ability to steer — a function that didn’t exist in previous rate proceedings — has created “workable competition” in the non-interactive webcasting market and called steering “price competition at work” in an “effectively competitive” market.

SoundExchange’s proposal of 0.25 cents per stream was based on the royalty rates of on-demand services (like Spotify) and rooted in a perceived “convergence” of features in on-demand and non-interactive services. It was firmly against using the steered rate in setting a benchmark. SoundExchange believed it to be unrepresentative of the larger market, and not something that would appear in real-world conditions. It further argued Pandora left out the value of non-statutory consideration that would raise the benchmark to “at worst, no lower than the compensation under the existing statutory rate paid by Pandora.”

As we said, what Merlin’s critics initially feared became reality. The judges used only the steered rate, not the headline rate, as one of the two benchmarks in setting the 2016 rate for free, non-interactive services. They weren’t compelled by Merlin’s and SoundExchange’s arguments that steering is not representative of actual market conditions and competition. Nor did they believe the steered and headline rates were influenced by the existence of the “pureplay” rate (discussed below), the rate standalone webcasters like Pandora were paying through 2015 (broadcasters like iHeartRadio paid the full statutory rate). These findings may seem counterintuitive, but much of the determination could be called counterintuitive.

The Judges’ reasons for choosing the steered rate are complex and steeped in economic theory. (Judge Strickler, an attorney with experience mixing economics and legal theory, has a masters in economics and currently teaches economics at Brookdale College). In short, they found that the steered rates were directly negotiated between a willing buyer and a willing seller in a competitive market. Steering is “synonymous with price competition in this market,” the judges wrote, adding “competing sellers don’t engage in a race to the bottom,” they create a race to “a workably and effectively competitive price.”

In explaining their disagreement with SoundExchange, the judges cited testimony from Darius Van Arman, co-founder and co-owner of indie record label Secretly Group, in which he said it was in his label’s self-interest to act defensively and enter Merlin’s agreement with Pandora, since other labels could also reach agreements with the service. This exemplifies the price competition the judges found in steering. If Van Arman didn’t take a lower price for steered streams, his competitors might have.

Merlin opposes the choice of the benchmark. The judges “fundamentally misread the context of our agreement — and the fact it was negotiated subject to the narrow confines and availability of the Pureplay Settlement rates,” said Merlin CEO Charles Caldas in a statement given to Billboard. “Given these constrictions, we fail to understand how the results could be interpreted as a benchmark for a true free market rate.”

There is another issue here. Merlin, SoundExchange and the services believe both the headline rate and the steered rate reside in the “shadow” of the statutory rate. That is to say, the existing statutory rate acts as a starting point for any direct negotiations. A buyer (like Pandora) could always fall back on the statutory rate if the seller (the label) asks too high a price. The “shadow” argument says a negotiated rate would be different if there were no government-set, statutory rate in the first place. A second shadow is the statutory rate yet to be set.

But the judges didn’t find a compelling shadow effect. Writing about the Merlin/Pandora and WMG/iHeart deals, the judges said the sellers (labels) were not affected “because… they voluntarily agreed to rates below the applicable statutory rates (in exchange for the steering of more plays), rather than defaulting to the higher statutory rate.” Again, this is counterintuitive if you believe the headline rate could be influenced by the mere existence of a statutory rate. This finding would surprise Merlin labels. According to a source with knowledge of the Merlin/Pandora agreement, the headline rate is pegged to the Pureplay rate through 2015 and the statutory rate in 2016 and beyond.

It’s easy to imagine a chilling effect on future direct agreements. A label has little incentive to perform direct licensing deals with services if a headline or steered rate falls below the statutory rate. A service willing to trade non-financial benefits for a lower rate would have a more difficult time negotiating with a label. But the judges weren’t about to tear down the process and start over. If there were a shadow effect, “[t]hat could mean the wholesale of abandonment of benchmarking, to be replaced by a valuation approach yet to be applied and accepted in these proceedings,” they wrote.

Another important factor in setting a benchmark is making an adjustment to account for additional items (both financial and non-financial) in the agreement. For a variety of reasons, Pandora expert witness Dr. Shapiro did not assign separate values to terms outside the statutory license. The labels also did not provide values for additional items (such as advertising placements), and its witnesses tried to downplay the value of financial and non-financial elements in the agreement. Merlin’s Charlie Lexton stated his organization never took the promotional and substitutional benefits into account when reaching an agreement with Pandora. SoundExchange insisted there was “no precise calibration” of these elements before the negotiations.

The judges expressed surprise. “It strains credulity to think Merlin was oblivious to the potential promotional and substitutional effects of the Pandora/Merlin Agreement, yet still proceeded with the deal on unaltered terms,” the judges wrote in response. They ended up deciding “Merlin found the value in the Pandora/Merlin Agreement to lie in the… trade-off of more plays at a lower rate for more total revenue” and declined to adjust the benchmark for any “extra-statutory consideration.”

Merlin disagrees with the judges, saying in a statement to Billboard that “the judges have appeared to selectively disregard key benefits of the partnership, which were negotiated in addition to per play royalties.”

The other benchmark for setting the free, non-interactive rate was WMG’s agreement with iHeartRadio. The judges found the stated or “average” rate, not the steered rate, “to be a useful benchmark” that satisfies the willing buyer-willing seller test.

The WMG/iHeart agreement also contained considerations that might merit adjustments to the benchmark rate. The judges made clear they expected SoundExchange and iHeartRadio to present this evidence to rebut the testimony of their adversaries “unless no such evidence — factual or expert — could reasonably be presented.” After all, it would be in each party’s interest to influence the benchmark’s adjustment in their favor.

Yet again, the judges were not presented any internal valuations. Although this portion of the decision is heavily redacted, the judges made clear neither WMG, SoundExchange nor iHeart placed a value on an element of the contract they believed to merit an adjustment of the benchmark. And again, the judges exhibited surprise about the “chronic problem” that none of the parties presented evidence which could place a value on an element that could help their case.

Everything up to this point involved the free, non-interactive rate. But the judges also set a rate for subscription, non-interactive services. It was here that SoundExchange’s “convergence” argument, made in regards to the free rate, made an impact. “Convergence” in this context means increasing similarity between the features of on-demand services like Spotify and non-interactive services like Pandora that has “blurred” the similarity between the two. On-demand services have become more “lean back,” meaning more radio-like, while noninteractive services offer more “lean-forward” with better personalization and features like “Pandora Premiers,” an on-demand selection of new releases.

While the judges found “significant evidence of functional convergence (up to the limits of the DMCA),” they decided convergence is relevant “only in the subscription market.” And so the set the subscription benchmarks via those proposed by SoundExchange’s expert witness, Dr. Daniel Rubinfeld, for the free benchmark. One benchmark was SoundExchange’s rate for non-interactive services that Dr. Rubinfeld adjusted downward by 12 percent to 0.21 cents. The second benchmark, unknown due to heavy redactions, established “an extremely tight zone of reasonableness,” according to the judges. The subscription rate determination ended up being 0.22 cents.

SoundExchange declined to comment and referred Billboard to the statement in made after the judges released their decision in December. Like Merlin, it also believed the judges’ decision did not reflect actual market conditions. “Our rate proposal used data from dozens of marketplace deals and was based on what willing buyers and sellers would agree to,” it said.

Pandora expressed contentment when the ruling was announced in December. Its effective rate, a weighted combination of the free and subscription rate, is 0.176 in 2016. “This is a balanced rate that we can work with and grow from. The new rate structure will enable continued investment by Pandora to drive forward a thriving and vibrant future for music,” CEO Brian McAndrews said in a statement.

There are always post-rate proceeding appeals. Merlin has pledged to “do what we can to correct this situation.” An appeal of a CRB decision is standard but has a low chance of success. After the Register of Copyright completes a review and the decision is officially published in the Federal Register, the parties have 30 days to file a notice of appeal to the District of Columbia Court. SoundExchange’s motion for rehearing has been denied. The CRB made available its response on Wednesday.